Automatic Avalanche, Landslide and Rockfall Detection

The reliable detection of snow avalanches, landslides and rockfall provides the basis for any advanced investigation on triggering mechanisms, risk assessment or the identification of possible precursors. It is well known that such phenomena produce specific seismic signals. Nonetheless, the manual detection on seismic traces is not feasible since it is extremely time consuming and results are influenced by the subjective view of the analyst. Automatic detection is therefore preferred, as unbiased results are obtained in near real-time. In this project, we will take advantage of an automatic classification procedure for continuous seismic signals to detect avalanches, landslides and rockfalls in continuous streams of seismic data. The applied method is independent of previously acquired data and classification schemes, offering the opportunity to detect very rare and highly variable events.

The goal of this project is to improve the detection across the entire Swiss Alps and to better understand possible triggering mechanisms. In addition accurate information on the number and release time of avalanches will provide an important contribution to the avalanche warning service at the SLF. This research is highly relevant since large scale avalanche and landslide monitoring will provide reliable data for further research, the development of near real-time products and improved risk assessment and warning systems.